The Future is Now: How Facial Recognition, AI, NLP, and Genetic Testing Aid in Detecting Rare Diseases

The Future is Now: How Facial Recognition, AI, NLP, and Genetic Testing Aid in Detecting Rare Diseases

Rare diseases, though often overlooked, have profound impacts on those affected. 

Out of over 10,000 rare diseases identified, 75% affect children, with one in 12 babies born with such conditions. Tragically, nearly 30% of these children do not survive beyond five years. 

While diagnosing rare diseases presents significant challenges, advancements in technology - such as Natural Language Processing (NLP), Artificial Intelligence (AI), and Genetic Testing - offer promising support to healthcare professionals, helping them detect, diagnose, and treat these conditions more effectively.

The Role of Technology in Rare Disease Detection

1. Natural Language Processing (NLP)

NLP technology is changing how medical data is analyzed, enabling faster and more accurate diagnosis of rare diseases. Traditionally, reviewing a patient's medical history involves sifting through vast amounts of data, a time-consuming task requiring human intervention. 

NLP software simplifies this process by analyzing structured and unstructured data, extracting key details, and providing healthcare professionals with a comprehensive view of a patient’s condition. This approach significantly shortens the timeline for diagnosis, which is crucial for rare disease patients.

Example: University of Iowa Stead Family Children’s Hospital

• Researchers at this hospital utilized NLP to streamline the detection of rare diseases. By applying NLP to Electronic Health Records (EHRs), they were able to quickly and accurately identify high-quality phenotype information from patients who had undergone genomic testing.

• Benefits: This method increased the speed of phenotype analysis by 200 times compared to manual processes. By leveraging this data, healthcare professionals can more efficiently diagnose rare diseases, improve patient records, and identify suitable candidates for clinical trials.

Example: Shire Pharmaceuticals and Hunter Syndrome

• Shire used Natural Language Processing text mining to identify patients for a clinical trial aimed at treating Hunter Syndrome, a rare condition that typically manifests in early childhood and progresses rapidly, often leading to death by the second decade of life.

• Benefits: The NLP-driven search of scientific abstracts enabled precise identification of patients who would benefit most from the therapy, improving the chances of successful treatment and highlighting NLP’s potential in addressing rare diseases.

2. Artificial Intelligence and Deep Learning

AI and Deep Learning are making significant strides in the medical field, particularly in diagnosing rare diseases. These technologies can analyze complex data, such as facial photographs, to identify health issues that may be difficult to detect through traditional methods.

Example: Deep Learning and Acute Illness Detection

• A study in 2021 investigated whether a Deep Learning algorithm could distinguish between healthy individuals and those acutely ill based solely on facial images. The study was based on the clinical gestalt theory, which posits that analyzing facial and body cues can enhance health assessments.

• Benefits: The algorithm successfully identified acutely ill patients, suggesting that AI could assist in diagnosing conditions like sepsis in emergency settings, either on its own or in combination with other diagnostic tools.

3. Facial Recognition Technology

Facial recognition technology has proven valuable in detecting rare diseases, especially those with visible physical manifestations. This technology uses AI algorithms to analyze facial features and compare them against a database of known conditions, providing healthcare professionals with diagnostic suggestions.

Example: Face2Gene by FDNA

• Face2Gene, an app developed by FDNA, uses facial recognition to detect rare diseases in children. It analyzes a single facial photograph and suggests the top ten possible conditions, with a 91% accuracy rate. However, it does not provide a definitive diagnosis, and further testing is required.

• Benefits: The app has become an essential tool for healthcare professionals, particularly during the COVID-19 pandemic, as telemedicine adoption has surged. While it cannot detect diseases that do not affect physical appearance or purely neurodegenerative conditions, it remains a powerful resource for early diagnosis.

Example: 22q11.2 Deletion Syndrome Detection

• Researchers at the National Human Genome Research Institute used facial recognition software to detect 22q11.2 deletion syndrome, a rare disease affecting 1 in 3,000 to 1 in 6,000 children, across various ethnic groups.

• Benefits: The software achieved a 96.6% diagnostic accuracy across different ethnicities, demonstrating its potential to assist clinicians in diagnosing rare diseases quickly and accurately, regardless of the patient’s background.

Genetic Testing and Phenotyping

Genetic testing and phenotyping are critical in identifying rare diseases, particularly those with genetic origins. While genetic testing analyzes a patient’s DNA for abnormalities, phenotyping examines observable traits resulting from the interaction of genes and the environment.

Example: Combining Phenotyping and Genotyping

• Phenotyping and genotyping complement each other in detecting rare diseases. While phenotyping suggests which genes to investigate, genotyping pinpoints the affected genes. Together, these methods enhance the accuracy of rare disease diagnosis, benefiting the over 300 million people worldwide affected by these conditions.

Guest Take: Misty Milby, VP Clinical Business Development at Everly Health